The goal of this project is the development of new antineoplastic agents directed against tubulin, a protein critical for cell division and form the maintenance of cellular morphology. Detailed mechanistic and structure-activity studies were performed with many classes of compounds. Analogs of combretastatin A-4 were examined to determine important structure-activity relationships, and studies to determine the optimal bridge length between phenyl rings were completed. The interaction with tubulin of the pentapeptide dolastatin 10, a noncompetitive inhibitor of vinca alkaloid and nucleotide binding, was defined with great precision based on properties of stereoisomers and segments of the active molecule. Mechanistic studies were initiated with dolastatins 11, 14 and 15, compounds which do not appear to interact strongly with tubulin in vitro but which disrupt the function of cellular microtubules. Structure-activity and mechanistic studies were continued or initiated with analogs of colchicine, 2-methoxy-5- (2',3,'4'-trimethoxyphenyl)tropone, derivatives of 2-styrylquinazolin-4 -one (SQZ), and derivatives of 5,6-diphenylpyridazin-3-one (DPP). The SQZ derivatives were defined as compounds binding rapidly and reversibly, but with low affinity, at the colchicine site. The DDP derivatives appear to bind at a unique binding site on tubulin. A radiolabeled A-ring modified analog of colchicine was synthesized. It reacts covalently with beta-tubulin. We initiated studies with photoactive and chemically reactive analogs of taxol which should ultimately prove useful in defining the taxol binding site on tubulin. The DTP drug screening data base was examined using the COMPARE program with known antimitotic compounds as seeds to yield multiple leads for new antimitotic agents. The most potent were halichondrin B and homohalichondrin B. These agents were shown to be noncompetitive inhibitors of vinblastine binding to tubulin.